The Value of Political Connections in Fascist Italy — Stock Market Returns and Corporate Networks ✩ Tiziana Foresti 1 , Nadia Garbellini 2 , Ariel Luis Wirkierman 3 Keywords: Fascism, Event Studies, Social Network Analysis, Italian business history 1. Introduction Recent years have witnessed the flourishing of a body of economic literature concerned with the search for empirical evidence of a positive relation between political connections, economic rent and the value of firms. 4 The present paper contributes to the strand of this literature that deals with the quantitative measurement of the value of the political connections of firms. 5 Our work proposes a quantitative measurement of the value of political connections between Italian firms and the Fascist regime in the years of Benito Mussolini’s rise to power (1921-1925). Specifically, the present paper offers a quantitative answer to the question: how much was it worth to have close, early connections with the National Fascist Party (hereafter, PNF)? Following in the tracks of Ferguson and Voth (2008), who studied the reaction of the German stock market to the Nazi seizure of power, in order to assess the value of political connections we perform an event study, which consists of an estimation of the impact of unexpected events on stock market returns (MacKinlay, 1997; Campbell et al., 1997). In particular, the present paper analyzes the reaction of Italian stock market investors to the March on Rome, the Fascist military expedition of 28 th October 1922 with which the first Mussolini government unexpectedly began. ✩ Research output corresponding to Inet Grant No. INO1400003. This paper is still work in progress, hence this version is preliminary, please do not cite without authors’ permission. Comments, critiques and suggestions are kindly welcomed. Email addresses: [email protected](Tiziana Foresti), [email protected](Nadia Garbellini), [email protected](Ariel Luis Wirkierman) 1 Baffi Carefin, Centre for Applied Research on International Markets, Banking, Finance and Regulation, Bocconi University, Milan, Italy. 2 Dipartimento di Ingegneria Gestionale, Universit` a degli Studi di Bergamo, Italy. 3 Dipartimento di Discipline matematiche, Finanza matematica ed Econometria, Universit` a Cattolica del Sacro Cuore, Milan, Italy. 4 See, for example, Shleifer and Vishny (1994); Khwaja and Mian (2005); Faccio et al. (2006); Bunkan- wanicha and Wiwattantang (2009); Boubakri et al. (2012). 5 See, for example, Roberts (1990); Agrawal and Knoeber (2001); Fisman (2001); Johnson and Mitton (2003); Li et al. (2008); Civilize et al. (2015). Preprint submitted to INET VI Annual Conference, April, 8-11, 2015, Paris April 6, 2015
Transcript
The Value of Political Connections in Fascist Italy — Stock Market
Returns and Corporate NetworksI
Tiziana Foresti1, Nadia Garbellini2, Ariel Luis Wirkierman3
Keywords: Fascism, Event Studies, Social Network Analysis, Italian business history
1. Introduction
Recent years have witnessed the flourishing of a body of economic literature concernedwith the search for empirical evidence of a positive relation between political connections,economic rent and the value of firms.4 The present paper contributes to the strand ofthis literature that deals with the quantitative measurement of the value of the politicalconnections of firms.5
Our work proposes a quantitative measurement of the value of political connectionsbetween Italian firms and the Fascist regime in the years of Benito Mussolini’s rise to power(1921-1925). Specifically, the present paper offers a quantitative answer to the question:how much was it worth to have close, early connections with the National Fascist Party(hereafter, PNF)?
Following in the tracks of Ferguson and Voth (2008), who studied the reaction of theGerman stock market to the Nazi seizure of power, in order to assess the value of politicalconnections we perform an event study, which consists of an estimation of the impact ofunexpected events on stock market returns (MacKinlay, 1997; Campbell et al., 1997). Inparticular, the present paper analyzes the reaction of Italian stock market investors to theMarch on Rome, the Fascist military expedition of 28th October 1922 with which the firstMussolini government unexpectedly began.
IResearch output corresponding to Inet Grant No. INO1400003. This paper is still work in progress,hence this version is preliminary, please do not cite without authors’ permission. Comments, critiques andsuggestions are kindly welcomed.
1Baffi Carefin, Centre for Applied Research on International Markets, Banking, Finance and Regulation,Bocconi University, Milan, Italy.
2Dipartimento di Ingegneria Gestionale, Universita degli Studi di Bergamo, Italy.3Dipartimento di Discipline matematiche, Finanza matematica ed Econometria, Universita Cattolica del
Sacro Cuore, Milan, Italy.4See, for example, Shleifer and Vishny (1994); Khwaja and Mian (2005); Faccio et al. (2006); Bunkan-
wanicha and Wiwattantang (2009); Boubakri et al. (2012).5See, for example, Roberts (1990); Agrawal and Knoeber (2001); Fisman (2001); Johnson and Mitton
(2003); Li et al. (2008); Civilize et al. (2015).
Preprint submitted to INET VI Annual Conference, April, 8-11, 2015, Paris April 6, 2015
In the years of Mussolini’s rise to power, the Italian stock market was less advanced than,for example, the German one. To begin with, in Italy there were seven stock exchanges:Milan, Rome, Turin, Genoa, Trieste, Naples and Florence. There was a sort of topographicdivision among the different markets that reflected the industrial specialization of eachregion. For example, Milan was the leading Italian stock market for state bonds, bankingbonds, and many others securities. Turin was the leading stock market for electric stocks.Genoa was the leading stock market for nautical stocks and sugar stocks. Secondly, in 1921only 120 societies were quoted in Milan, which was the leading market for exchange value(Consob, 2011, chapters 1 and 2). Thirdly, stock-market transactions were often somewhatopaque because of a structural conflict of interests due to the connection between listedcompanies and ‘mixed’ banks.6 In other words, a few big banks and the group of firms thatwas financed by them tended to form something like credit-industrial organizations that,mutually, influenced the conduct of business and credit management (Bonelli, 1971). Aswe shall show, our innovative analysis not only unveils the clustering configuration in termsof political connections with PNF of Italian industry during the interwar period, but alsoshows that the firms in the connected cluster outperformed the rest of the economy in termsof stock market returns.7
Mussolini’s assumption of power in November 1922 was facilitated by a complex concur-rence of economic and social factors. At the end of the World War One, Italy went througha deep economic crisis. War expenses had been supported by government borrowing. OnDecember 19th 1914, on June 15th 1915, and on December 22nd 1915 the government autho-rized three ‘National Loans’ (Prestiti Nazionali). These loans, redeemable in 1925, yielded4.5% interest in the case of the first two and 5% in the case of the third. On January 2nd
1917, on December 6th 1917, and on November 24th 1919 the government issued consolidateddebts at 5% interest. Northern Italy, home to only 46% of the population, absorbed 60%of the loans. Moreover, Italy had heavy international debts — owing 611 million poundsto Great Britain and 1648 million dollars to the United States (Ministero del Tesoro, 1999,chapter 3).
The war-profits tax of 1915 and the capital levy taxes of 1919 and 1920 notwithstanding,by the end of the war Italy faced a persistent budget deficit and a predominant short-term floating debt in the structure of the public debt (on June 30th 1921, for example, itamounted to 36.5% of the national debt; Ibid.: 35). The post-war economic situation wasfurther worsened by the inflationary trend triggered by a massive note issue in the yearsbefore.
In addition, the numerous governments that followed each other until 1922 had to facea difficult social situation characterized by repeated strikes in cities and country. The socialdiscontent arose out of increases in the prices of basic goods, such as bread, and a highrate of unemployment due to the industrial reconversion of the factories involved in military
6At the beginning of the 20th century, the Italian bank system was a peculiar mix of traits from theAnglo-Saxon market-oriented model and the German model of the universal bank.
7For a general description of the characteristics of Italian industry during the interwar period see Amatoriand Brioschi (1997); Amatori and Colli (1999); Amatori et al. (2011).
2
production and the mass of ex-servicemen.This is the context in which Mussolini, following his interventionist position with regard
to the war, was able to exploit the expectations of ex-servicemen and to present the PNFto the public as a safeguard of public order by means of the fascist action squads. At thesame time, he cemented his relationships with the big magnates of Italian industry thanks tothe anti-union bias that, from 1920, increasingly characterized the fascist movement. As weshall see, some of the big names of electrical and metallurgic industry financed Mussolini’spolitical project in its initial stage. In this respect, our work enriches current knowledgeconcerning the origin of the flow of capital that financed Mussolini in these years.
As for the Italian Stock Market, the years after World War One were characterized bystrong financial instability, which was worsened in the early 1920s by the difficulties of twoimportant banks: Banca Italiana di Sconto and Banco di Roma.8 Between 1918 and 1922the real stock return decreased 25% (Consob, 2011, p. 20). In the years from 1923 to 1925,that is during the first Mussolini government, shares prices were boosted by the economicrecovery of those sectors which were more open to exports, such as textiles and machinery,and those which were more innovatory, such as the electrical sector.9
Nevertheless, conclusive evidence of the sizeable impact of political connectedness onstock market performance suggests the significant involvement of big business in politicaldevelopments during a key period of modern history. Moreover, such a finding signals theneed to perform a more careful reading of business history (Lamoreaux et al., 2007). Specif-ically, the existence of a measurable advantage for connected firms suggests that key playersbehind the widening of Italy’s industrial structure were not only driven by effective demandfor their output, but also by the differential returns to be obtained from political support ofFascism. Indeed, the possibility for the firms in the connected cluster to outperform the restof the economy in terms of stock market returns calls into question studies in firm dynamicsthat exclude the role of power and politics in the capitalist competition process (Ferguson,1995).
The rest of the paper is organized as follows. Section 2 offers a general description of thephases of Mussolini’s rise to power. Section 3 focuses on the extent of Italian magnates’ sup-port for Mussolini. Section 4 explains the characteristics of our dataset. Section 5 presentsthe results of our research into firms connected with the PNF by means of a network analy-sis. Section 6 illustrates the methodology of an event analysis. Section 7 presents the resultsof the employment of this technique in the case of the Milan Stock Exchange during theMarch on Rome. Appendix A contains a historiographic analysis of prominent individualswithin the cluster of connected firms. Appendix B provides a statistical companion.
2. Mussolini’s rise to power
As is well known, in spring 1921 Mussolini transformed the Fascist movement into thePNF and, in the general elections that took place in May 1921, 32 candidates of the PNF,
8On this bank crisis see Asso and De Cecco (1994).9The increase in exports during the first Mussolini government was due to the positive international
economic climate as well as depreciation of the lira exchange rate (De Cecco, 1993).
3
together with Mussolini, became deputies.In the following months Mussolini defined the political platform of the PNF and, no
doubt with an eye to following the parliamentary path to power, agreed with the SocialistParty to end the violence of the fascist action squads (Patto di Pacificazione). The strongopposition of the fascist actions squads towards this Party line, however, soon compelledMussolini to break his agreement with the Socialist Party. Summer 1921 saw a new period ofsocial riots, acts of violence against political opponents, and the progressive transformationof the fascist action squads into a paramilitary organization.
The weakness of the Bonomi Government (July 4th 1921 — February 26th 1922) and ofthe two Facta Governments (respectively, February 26th 1922 — August 1st 1922 and August1st 1922 — October 31st 1922) in opposing the unlawful actions of the fascist action squads,together with the inability of the leaders of the other Italian political parties to overcometheir individual interests in order to establish a joint front against Fascism, paved the wayto Mussolini’s rise to power.
In this context of political instability, on August 1st 1922 the trade unions called anew general strike in support of workers’ rights. This strike again gave to the Fascists thepossibility of presenting themselves before the public as the only Party able to maintainpublic order. In many cities, such as for example Milan, members of the PNF assumedresponsibility for providing public transport. However, the strike also gave the fascist actionsquads the opportunity to attack the seats of trade-union organizations across the country.The PNF was thus simultaneously maintaining public order and committing illegal acts; butit was the former that primarily attracted public attention. Because of the strikes 7.336.393workdays were wasted in Italy between November 1st 1921 and October 31st 1922; of these6.892.795 were workdays of the manufacturing sector (De Felice, 1966a, vol. I: 396).
The March on Rome (Marcia su Roma) of October 1922 was, then, the culmination ofthe long political and social crisis that began after the general elections of May 1921. In thebuild up to this Fascist military expedition to Rome three dates have a crucial importancefor us: the 16th, 24th and 28th of October 1922 (Vivarelli, 1992, vol. III: 435-454).
In Milan, on 16th October 1922, Mussolini met with some generals of the Italian armywho were also members of the PNF in order to seek the support of parts of the Italian army.In Naples, on 24th October 1922, during a secret meeting with his closest collaborators,Mussolini fixed the date of the March on Rome: that is, October, 28th 1922.
The revolt was to be carried out by means of two different kinds of operation. Theadvance on Rome would be concurrent with the taking of the Prefectures and the editorialoffices of the most important newspapers in many Italian cities. Nevertheless, and as weknow, none of the Fascists Marched on Rome until King Vittorio Emanuele III of Savoiagave Mussolini (who was in Milan) the task of forming a new government on October, 29th
1922. On the morning of October 28th, the King refused to sign the decree that had beendrawn up by the Government during the night. The Prime Minister, Facta, asked the Kingto declare a state of siege in order to allow the army to defend Rome. The army undoubtedlyhad an advantage over the fascist action squads, being better equipped and better drilled,and so today it is hard to understand why the King did not sign the decree on the state ofsiege and, in fact, compelled Facta to resign.
4
On November 16th 1922, Mussolini called the Chamber of Deputies for a vote of confi-dence on his first government and, on November 29th, he won a vote of confidence at theSenate.
Mussolini had the offices of both Foreign Minister and Home Minister. Three othermembers of the PNF were appointed ministers: Alberto De Stefani became Minister of Fi-nance, Aldo Oviglio was given the office of Minister of Justice, and Giovanni Giurati wasput in charge of the ministry that administered the territories annexed to Italy after theFirst World War (Ministero per le Terre liberate dal Nemico). The other ministers weremembers of the Popular Party, the Liberal Party and the Democratic Party. High AdmiralPaolo Emilio Thaon di Revel was appointed Minister of Navy and Armando Diaz, who onNovember 8th 1917 had been nominated Chief of Staff by the King, held the position of Min-ister of Defence.10 Nine out of eighteen vice-ministers were fascists. Thus the first Mussolinigovernment undoubtedly had a fascist complexion, but was nevertheless the expression of aparliamentary coalition (Ibid.: 480).
However, as leader of the PNF, Mussolini had to face the problem of the normalizationof the fascist action squads. With this in mind, the government established the Miliziavolontaria per la sicurezza nazionale (hereafter, MVSN) with a royal decree of January1923. Because the MVSN was under the authority of the Prime Minister, the fascist actionsquads were removed from the control of the local members of the PNF. Moreover, Mussolinisucceeded in legalizing the actions of the fascist squads because the purpose of the MVSN wasto defend the Fascist revolution and public order in collaboration with the army (De Felice,1966a, vol. I: chapter 5).
In concomitance with the establishment of the MSVN as a politicized security force,Mussolini gave up on passing any reform bill regarding military regulations and, notwith-standing the different view of the Finance Minister, De Stefani, he increased the nationalbudget for military expenditure. In the financial year 1921-1922 the army received a budgetof 1921 million lire rather than the budget that had been proposed for this year, that is 1571million lire. These complications regarding the army notwithstanding, on July 28th 1923 theMinister Diaz hailed the “highly patriotic goals of the MSVN” (Ibid.: 435).
In the first months of his government Mussolini, with an eye to calling general elections,partly amended the existing proportional electoral system with the so-called Acerbo law,which was passed by the Chamber of Deputies on July 23rd 1923 and by the Senate onNovember 13th 1923. By the Acerbo law, if a Party list reached a percentage of 25% ofvotes, that Party automatically received 2/3 of the total parliamentary seats. Moreover,the vote count had to be conducted on a national basis. Of 535 deputies, 365 would enterthrough this majority premium, the remaining 179 by means of the proportional system.
In April 1924 the PNF won the general elections with 65% of the votes. As is well known,Giacomo Matteotti, before his assassination on June 10th 1924 by a group of fascists, publiclydenounced the widespread electoral fraud of the 1924 general elections.11
On January 3rd 1925, Mussolini spoke to the Chamber of Deputies and took full political,
10At that time the Ministry of Defence was named Ministero per la Guerra.11On Matteotti’s assassination see Canali (2004, 2009).
5
moral and historical responsibility for Matteotti’s assassination; but he denied having beenthe instigator of the assassination. Mussolini, in fact, challenged the Parliament to impeachhim. Not only did Mussolini not suffer impeachment, but his speech sanctioned the trans-formation of Italy into a totalitarian state. In the following years the Italian Parliamentwas progressively deprived of its prerogatives and the general elections, held on March 24th
1929, took the form of a referendum.
3. Fascism and Industry
The extent of Italian magnates’ support for Mussolini is a controversial point. Broadly,scholars such as Rossi (1955), Guerin (1956) and Sarti (1977) explicitly link Mussolini’s riseto power with the attempt of the big landowners and magnates to check union demands inthe farmlands and in the factories. More cautiously, Melograni (1972) denies the existenceof any organic support to Mussolini from the big magnates, arguing that among them therewas a plurality of attitudes towards fascism.
As we shall see, Mussolini benefited from the financial support of some of the big namesin the electrical, the tyre, the steel and the iron industries in the form of financing of hisnewspapers and of the election campaign of the PNF in 1922. Moreover, Mussolini’s firstgovernment had the official confidence of both Confindustria (the Confederation of ItalianIndustry) and Assonime (the Association of Italian joint-stock companies), notwithstandingany unfavourable attitudes of some of their members towards fascism. On November 1st 1922Confindustria, in fact, claimed “to have exerted a direct and pressing influence in favour ofMussolini’s solution” (Rossi, 1955, p. 41).
In beginning the exploration of the complex relation between fascism and industry, itis necessary as a first step to outline briefly the structure of Confindustria and Assonimeduring the years of the birth of the fascist movement and of the success of Mussolini.
Following the experience of local entrepreneurial associations in Milan, Genoa and Biellabetween 1902 and 1906, and the establishment of the Lega Industriale of Turin, Confindustriawas founded on May 5th 1910 with the goal of coordinating at the national level the initiativesof entrepreneurs in their relations with the trades union and both the central and localgovernments. The founder of this employers’ association was Gino Olivetti, the generalsecretary of the Lega Industriale, and an expert of industrial organization. He was thegeneral director of Confindustria until 1934.
Confidustria’s first president was Luigi Bennefon Craponne, a French silk industrialistwho was in office until 1913. In the years 1914-1918 he was followed by Ferdinando Bocca,the head of a leather tanning industry. In 1919 Confindustria had two presidents, first DanteFerraris, who resigned because he was nominated Minister of Industry, Commerce and Work(Ministro dell’industria, commercio e lavoro) in the Nitti government, and then GiovanniBattista Pirelli, who headed the most important Italian tyre factory. Giovanni Silvestri tookoffice as president in the years 1919-1920. With his Officine Meccaniche Miani-Silvestri, heis considered one of the pioneers of the Italian mechanics industry. Silvestri was followedby Ettore Conti (1920-1921), who boosted the exploitation of hydraulic force for producingelectric energy in Northern Italy. From 1922 to 1923 the president was Raimondo Targetti,
6
who operated in the wool industry. Targetti was succeeded by Antonio Stefano Benni, whostayed on until 1934. Benni headed Fabbrica Italiana Magneti Marelli, a factory of electricalmachinery.
Assonime was founded on November 22nd 1910 on the initiative of 53 businessmen onbehalf of 181 companies. Their objective was to reduce the sphere of state interventionin the economy and to obtain tax cuts for joint-stock companies. The following year 503companies joined Assonime. Carlo Esterle, one of the pioneers of the electrical industry inItaly, was the president of Assonime from its foundation to 1917. Esterle was followed byFerraris, who held the position of president until 1919. Luigi Volpi di Misurata succeededFerraris. His main company was the electrical company Societa Adriatica di Elettricita.From 1921 to 1922 the president of Assonime was Silvestri. In the years 1922-1924 the officewas held by Conti, who was followed by Alberto Pirelli, son of Giovanni Battista, who heldthe position until 1945.
Confindustria was formed as an apolitical association. Olivetti, aiming to achieve con-sensus among the members of the association, decided that Confindustria did not supportexplicitly any political parties (Belloni, 2011, chapters 1 and 2). This resolution was soondisputed in 1911, when a government monopoly of life assurances was planned, and in 1913when the government threatened to expel president Craponne from Italy for public nuisancebecause he was embarking on a lockout of the automobile industry in Turin in reaction toworkers’ strikes. In 1915 Confindustria officially took an interventionist position with regardto the war. In the elections of 1919 Olivetti became deputy in the ranks of the right wingof the Liberal Party and, henceforth, effectively acted as a hinge between the industrial andthe political classes.
We have now covered the ground necessary to understand why Confindustria made astand for fascism immediately after the March on Rome.
The anti-union bias, which from 1920 increasingly characterized the fascist movement,undoubtedly appealed to Italian magnates. The wave of strikes that began in the springof 1919 as a reaction to the heavy economic crisis that beset Italy in the wake of WorldWar One, resulted in the occupation of metallurgic factories throughout Italy in September1920. Workers demanded not only a wage rise and an eight-hour working day, but alsothe establishment of workers committees in the factories. These committees, they insisted,should take active part in the company management.
The Giolitti government (June 16th, 1920 – July 4th, 1921) remained neutral and refrainedfrom calling out the police in order to free the factories (such as for example FIAT in Turin)that were occupied. Instead, Giolitti resolved to wait for the realization of a compromisesolution between workers and industrialists.
Olivetti set out Confindustria’s intransigent position against workers committees inMarch 1920, during a meeting of Confindustria delegates in Milan. Olivetti’s position wasgrounded on matters of principle and practice: the former in relation to the revolutionarybias of the proposal; the latter arose from the fact that the workers committees were to beself-governing bodies, so industrialists could not negotiate collective agreements, as in thecase of trade unions.
Probably, Confindustria’s intransigence was not free from political opportunism, with
7
the industrialists hoping for some government benefits in order to sweeten the search for acompromise solution (Vivarelli, 1992, vol. II. pp. 592-646). As we know, the first financialprovisions of the Giolitti government caused a widespread outcry in industrial and financialcircles. On September 24th 1922, the government passed laws according to which 1) thestate took upon itself the extra profits generated by the war; 2) a parliamentary committeewould conduct an inquiry into the war expenses; 3) the rate of probate duty would becomemore progressive; 4) the motor vehicle excise duty would increase; 5) it was obligatory toregister all financial bonds, with the exception of government securities.
With these measures, Giolitti aimed to better the conditions of the working classes inorder to pave the way to the abrogation of the ‘political price’ of bread. Nevertheless, hiseconomic policy had deflationary effects and no provision had been made about customspolicy and farm aid.
In such a strained political situation, the occupation of factories was not just a matter ofpublic order. In the end, the government encouraged a resolution of the industrial disputesthat favoured the workers. On September 20th 1922 the trades unions (Fiom and C.G.L.)and Confindustria drew up an agreement. The terms agreed granted an improvement in theeconomic and working conditions of the workers and, more importantly, the possibility ofintroducing “workers’ control” inside the factories.
Giolitti’s choice of neutrality had overlooked the fact that the workers’ demands followedin the wake of the Russian Revolution and, for industrialists, the establishment of somethinglike workers committees inside their factories could appear as the first step towards the abro-gation of private property. As we have already seen, in the general context of parliamentaryweakness, Mussolini took advantage of the industrialist’s perception of a communist threatand was able to present the PNF to the public as an opponent of union demands and as ameans of maintain public order by means of the fascist action squads.
In recent years, historical research has enriched our knowledge of the origin of the flow ofcapital that financed Mussolini’s political project in its initial stages. According to De Felice(1966b, p. 277), Il Popolo d’Italia, the newspaper founded in 1914 by Mussolini in Milan, wasfinanced by capital from some French political circles and also a group of Italian industrialistsinterested in Italy’s entry into the War. These Italian industrialists were Ferdinando MariaPerrone (his Ansaldo was an iron and steel business), Esterle (his Edison was an electricalindustry), Giovanni Agnelli, the founder of FIAT, Emilio Bruzzone (his Unione Zuccheriwas a sugar industry) and Vittorio Emanuele Parodi (the owner of shipping company). Oursurvey of the pages of Popolo d’Italia in search of the companies who bought advertisingspace, allows us to add Societa italiana per le lampade eletteriche Z (which was a companyof the Edison group), Pirelli, Officine Meccaniche Miani Silvestri, Migone (which was achemical industry) to the financing group of Mussolini. We remind the reader that in 1914Esterle was the president of Assonime, while Giovanni Battista Pirelli and Silvestri headedConfindustria in the years from 1919 to 1920.
Archival research has proved that Banca Commerciale Italiana (hereafter, COMIT) pro-vided official financial support to Mussolini’s newspaper (Fabre, 2003; Barbone, 2003).
The question of the extent of Italian magnates’ support for Mussolini is controversial;without any presumption of completeness, we would like to offer a few remarks. Probably
8
the initial support for Mussolini’s political project by some Italian magnates arose out ofthe possibility of an increase in profits due to the job order for the war. On the other hand,support of Mussolini during the days of the March on Rome appears to relate to the desirefor stable government. On October 28th 1922, a group of magnates that included AlbertoPirelli, Olivetti, Conti, and Benni met Mussolini in Milan. There are many different versionsof the meeting, but the official statement issued by Confindustria on November 1st 1922 isunquestionable.12
The financial policy of the first Mussolini government leaned toward laissez-faire. One ofthe first act of the government was to revoke the law of 1922 on financial bonds. The ministerDe Stefani simplified the taxation system and cut down the rate of duties on profits and onnew industrial constructions in order to foster the accumulation of capital. With the aimof reaching a break-even point, he pursued a policy of retrenchment in the administrationof the national railways and postal services. In November 1923, 65.000 temporary stateemployees were dismissed (Toniolo, 1980, chapter II).
However, the rescue of Ansaldo, Banco di Roma (a bank) and Ilva (a metallurgic business)did not fit the laissez-faire economic policy of the first Mussolini government. In the case ofAnsaldo, the government implemented the measures worked out by the Facta governmentand disbursed subsidies to Ansaldo. But the rescue of Banco di Roma by way of a handoutwas an initiative of the Mussolini government. In the case of Ilva, the rescue was made bytwo banks: COMIT and Credito Italiano. The government reduced the amount of the debtsof Ilva to the state for non-payment of taxes and some prepaid orders that had not beenmet.
The years of the first Mussolini government were characterized by rapid industrial growthgenerated by an increase in investment and exports, which was nevertheless realized bymeans of a policy of low wages that had the result of promoting labor intensive manufacturingand making Italy less competitive in the international context in the following years.
4. Dataset characteristics: Il Sole and IMITA.db
Our empirical analysis is based on daily stock prices from the Milan Stock Exchange(MSE, hereinafter) covering the period from July, 1922 to February, 1923. Being an explo-ration into the early twentieth century, we highly depend on local data sources. In particular,data comes from the financial newspaper ‘Il Sole: organo ufficiale della Camera di commercioe industria di Milano’.13 Besides daily price quotations, this specialised publication containsa weekly summary with additional data concerning, for example, the nominal value of eachsecurity, as well as the performance of Italian state bonds.
12See Rossi (1955) and Belloni (2011) for two different reconstructions of the meeting.13The daily newspaper ‘Il Sole’ was founded in 1865 on the initiative of a group of small entrepreneurs in
the textiles sector. In 1905, representatives of the machinery and banking sectors became shareholders inthe newspaper. From its beginning, one of the main functions of the newspaper was to acquaint dealers withtrends on the stock exchanges, in commodities markets, and the performance of companies and of marketprospects. For this reason, soon after its foundation ‘Il Sole’ became an official organ of the chamber ofcommerce of Milan.
9
Moreover, as regards complementary firm characteristics for those enterprises operatingin the MSE, like industry classification, regional origin, a synthetic balance sheet, as well asthe composition of the Board of Directors, we have used the ‘IMITA.db’ database.14 Thisdatabase contains a digitalization and codification of a series of Yearbooks published byCredito Italiano (1919, 1921, 1923, 1926) and Associazione fra le societa italiane per azioni(1928, 1937), including firm-level data for some benchmark years (e.g. 1921, 1927, 1936).15
Query capabilities allowed us to perform specific search operations by different criteria, anddownload one database record at a time. Hence, once the list of business firms operating inthe MSE has been defined, specific search and download operations have been performed.
Afterwards, both data sources (‘Il Sole’ and ‘IMITA.db’) have been merged into a uni-fied relational database, matching firms in the financial newspaper with those appearingin IMITA. Given the different labels employed by each data source, in some cases, thismatching operation has required to proceed with great care.16
5. Identification of connected firms
A pervasive feature of Italian capitalism, already present at the beginning of the 20th
century, consists in “the presence of a dense network of interests taking the form of frequentinterlocking relationships among the large firms”(Colli, 2006, p. 831). Thus, in order toidentify those firms operating in the MSE connected to the PNF, we have built and analyseda network with the structure of interlocking directorates.
Specifically, we proceeded as follows. First, by using the IMITA database, we haveidentified the people belonging to the Corporate Board of each firm in the dataset. Second,we computed a distance matrix whose elements are the number of people that each pairof firms had in common. Based on such a distane matrix, we implemented and refined acommunity detection algorithm known as ‘spectral bisection for modularity maximization’(developed by Newman, 2006a,b; Leicht and Newman, 2008) to identify firm clusters.
We identified ten such clusters, of which only one showed returns significantly higherthan the market model average.17
The list of connected firms, belonging to cluster 1 (CL1), is reported in Table 1; whilethe composition of non-connected clusters, from cluster 2 to cluster 10, is reported in Table2. Looking at the individuals enabling key connections among firms in each cluster, weidentified a set of people who actually had close connections to Mussolini.18
Thus, the logic of our empirical strategy may be termed as follows. On the basis of anunsupervised statistical learning algorithm we unconvered the community structure of the
14IMITA.db (IMprese ITAliane Data Base) has been created by a consortium of Universities (includingSiena, Bocconi, Bologna and Firenze), and supported by the Italian Ministry of Education, Universityand Research (MIUR) and the National Research Council (CNR). The database is freely accessible at:http://imitadb.unisi.it/en/home.asp.
15See Giannetti and Vasta (2006); Colli (2006); Colli and Vasta (2010) for details.16Useful information in this regard has been gathered from De Luca (2002).17See section 6 below for a detailed econometric specification of this statement.18See Appendix A for for a historiographic analysis of prominent individuals within the cluster of con-
nected firms.
10
graph of interlocking directorates of Italian firms operating in the MSE. Then, we studiedwhether any of these clusters (bond together thanks to individuals participating in differentcorporate boards) outperformed the market on the event of Mussolini’s rise to power. Asit turned out, only one cluster of firms had this feature. When studying the individuals atthe basis of this cluster, we could assert that they had indeed a connection to the Fascistenvironment. Thus, a machine learning algorithm together with historiographic analysis ofindividuals emerging from it allowed us to establish a link between political connections andstock market returns.
Figure 1 depicts the graph of the Italian MSE Corporate Network reflecting interlockingdirectorates.
Table 1: Clusters of Firms according to the shared members of their respective Corporate Boards
(Benchmark year: 1921; Cluster 1 identifies firms connected to the Fascist environment)Cluster of Connected Firms
Ticker Descriptor ATECO ATECO-Desc Cluster
Agr FonRust ISTITUTO DI FONDI RUSTICI SOCIETA AGRICOLA IN-DUSTRIALE ITALIANA
A01 Agriculture 1
FB Distillerie DISTILLERIE ITALIANE DA15 Food-beverages 1
FB IndZuc SOCIETA ITALIANA PER L’INDUSTRIA DELLO ZUC-CHERO INDIGENO
DA15 Food-beverages 1
FB MolAltaIt MOLINI ALTA ITALIA DA15 Food-beverages 1
FB Raffinerie SOCIETA LIGURE LOMBARDA PER LA RAFFI-NAZIONE DEGLI ZUCCHERI
DA15 Food-beverages 1
Cot Trobaso COTONIFICIO DI TROBASO DB17 Textiles 1Cot Turati COTONIFICIO FRANCESCO TURATI DB17 Textiles 1Cot Venez COTONIFICIO VENEZIANO DB17 Textiles 1Tess CascSeta FILATURA DEI CASCAMI DI SETA DB17 Textiles 1Tess UnManiff UNIONE MANIFATTURE DB17 Textiles 1Chi Bonelli FABBRICHE ITALIANE MATERIE COLORANTI
BONELLIDG24 Chemicals 1
Equip Ansaldo ITALIANA GIO. ANSALDO & C. DK29 Machinery-equipment 1
Elett Adriatica SADE SOCIETA ADRIATICA DI ELETTRICITA E40 Electricity-gas 1Elett UnEsEl UNES UNIONE ESERCIZI ELETTRICI E40 Electricity-gas 1Tran VeneteS VENETA PER COSTRUZIONE ED ESERCIZIO DI FER-
ROVIE SECONDARIE ITALIANEI60 Land-transport 1
Fin BdI BANCA D’ITALIA J65 Finance 1Fin BdRoma BANCO DI ROMA J65 Finance 1
Source: Own computations based on IMITA.db
6. An event-study on stock market returns
Once connectedness has been established, to measure its effect on stock market returnswe followed the event-study methodology (MacKinlay, 1997). Our event window coveredthe March on Rome.19 More precisely, a 21-day event window is employed, comprised of 10
19Following MacKinlay (1997), by event window we mean a brief time period associated to pieces of newswhich are supposed to influence stock market dynamics.
11
Table 2: Clusters of Firms according to the shared members of their respective Corporate Boards
(Benchmark year: 1921)Clusters of Other Firms
Ticker Descriptor ATECO ATECO-Desc Cluster
Min Elba ELBA SOCIETA ANONIMA DI MINIERE E DI ALTI FORNI CB13 Metal-ores 2Tess Targetti LANIFICIO TARGETTI DB17 Textiles 2Tess Tosi MANIFATTURA TOSI DB17 Textiles 2Tss ManDini MANIFATTURE TOSCANE RIUNITE DB17 Textiles 2
Chi Elettroc SOCIETA ITALIANA DI ELETTROCHIMICA DG24 Chemicals 2Chi Montecat MONTECATINI DG24 Chemicals 2Met Ilva ILVA ALTI FORNI E ACCIAIERIE D’ITALIA DJ27 Basic-metals 2TE Reggiane OFFICINE MECCANICHE ITALIANE DM35 Transport-equip 2
Elett Terni TERNI SOCIETA PER L’INDUSTRIA E L’ELETTRICITA E40 Electricity-gas 2
RT Rinasc LA RINASCENTE SOCIETA PER L’ESERCIZIO DI GRANDIMAGAZZINI
G52 Retail-trade 2
Tran FerMed SOCIETA ITALIANA PER LE STRADE FERRATE DELMEDITERRANEO
I60 Land-transport 2
Fin BCItal BANCA COMMERCIALE ITALIANA J65 Finance 2
Fin Brasital,Fin EspItalAm
SOCIETA PER L’ESPORTAZIONE E PER L’INDUSTRIA ITALOAMERICANA
J65 Finance 2
Cer RichGin CERAMICA RICHARD GINORI DI26 Non-metallic-mineral 3
Chi SNIA SNIA VISCOSA SOCIETA NAZIONALE INDUSTRIE APPLI-CAZIONI VISCOSA
DG24 Chemicals 4
Met Metalli METALLURGICA ITALIANA DJ27 Basic-metals 4MV FIAT F.I.A.T. DM34 Motor-vehicles 4TE MianiSilv OFFICINE MECCANICHE DM35 Transport-equip 4
Elett LigTosc SOCIETA LIGURE TOSCANA DI ELETTRICITA E40 Electricity-gas 4
Elett Vizzola VIZZOLA SOCIETA LOMBARDA PER DISTRIBUZIONE DI EN-ERGIA ELETTRICA
E40 Electricity-gas 4
Tran LibTries NAVIGAZIONE LIBERA TRIESTINA I61 Water-transport 4
Tcom Marconi SOCIETA ITALIANA SERVIZI RADIOTELEGRAFICI E RA-DIOTELEFONICI
I64 Post-Telecomm 4
Chi PirelliC PIRELLI & C. J65 Finance 4Fin CredItal CREDITO ITALIANO J65 Finance 4MV Bianchi FABBRICA AUTOMOBILI E VELOCIPEDI EDOARDO BIANCHI DM34 Motor-vehicles 5MV IsFrasc FABBRICA AUTOMOBILI ISOTTA FRASCHINI DM34 Motor-vehicles 5Elett Adamello GENERALE ELETTRICA DELL’ADAMELLO E40 Electricity-gas 5Elett Bresciana ELETTRICA BRESCIANA E40 Electricity-gas 5
Elett Conti SOCIETA ANONIMA PER IMPRESE ELETTRICHE CONTI E40 Electricity-gas 5
Elett Edison SOCIETA GENERALE ITALIANA EDISON DI ELETTRICITA E40 Electricity-gas 5
Elett Negri SOCIETA ELETTRICA RIVIERA DI PONENTE ING. R. NEGRI E40 Electricity-gas 5Tran NavGenIt NAVIGAZIONE GENERALE ITALIANA I61 Water-transport 5
Fin FerNaz SOCIETA ITALIANA PER LE STRADE FERRATE MERIDION-ALI
Met Camona OFFICINE DI SESTO SAN GIOVANNI & VALSECCHI ABRAMO DJ28 Fabricated-metals 7TE Breda ITALIANA ERNESTO BREDA PER COSTRUZIONI MECCA-
NICHEDM35 Transport-equip 7
Tess CanapNaz LINIFICIO E CANAPIFICIO NAZIONALE DB17 Textiles 8Man Mira FABBRICA CANDELE STEARICHE DI MIRA DN36 Manufacturing-nec 8Min Petroli PETROLI D’ITALIA CA11 Petroleum-Gas 9Tess Bernasc TESSITURE SERICHE BERNASCONI DB17 Textiles 9Cot ValOlon COTONIFICIO VAL D’OLONA OGNA CANDIANI DB17 Textiles 10
Source: Own computations based on IMITA.db
12
Figure 1: Graph of the Italian MSE Coroporate Network reflecting interlocking directorates(Benchmark year: 1921; Orange nodes represent firms belonging to the connected cluster; Nodes’ size reflects degree)
pre-event days, the event day (October 28th, 1922), and 10 post-event days.20
20Given that not all securities were traded in each day, the length of the pre-event and post-event days ofthe event window have been adjusted on a firm basis so as to cover the same period for all firms, whenever13
Moreover, as shown in Figure 2, each event window (T1, T2] has a prior estimation window(T0, T1], as well as a post-event window (T2, T3]. Their length is established in relation tothat of the event window, deciding also upon the level of granularity of stock market data(daily, weekly, monthly).21 In particular, our estimation window covers the three-monthperiod before the event window (i.e. July-October, 1922), and we have performed a weeklyanalysis based on daily data for stock prices.
Figure 2: Time line for an event study. Source: MacKinlay (1997, p. 20)
When the interest lies in testing an hypothesis on the relation between excess returns andfirm characteristics, the logic of event studies consists in a three-step procedure: (i) estimateexpected (or normal) returns without conditioning on the event taking place (i.e. duringthe estimation window), (ii) use these point estimates to linearly project abnormal returns(i.e. the difference between actual and estimated normal returns) during the event window,and (iii) adopt a cross-sectional regression approach to study how much of the variation inlog-returns is explained by specific firm characteristics like, for example, being politicallyconnected to the Fascist regime. A statistically significant coefficient associated to thisfeature would measure the value of political connections.
As regards step (i), we have estimated for each security expected returns by means of themarket model (MacKinlay, 1997, p. 18).22 As to step (ii), we computed cumulated abnormalreturns (CAR) during the 21-day event window for each security and, finally, as regardsstep (iii), we estimated the conditional expectation of stock market returns, conditioned tobelonging to the cluster of connected firms, as well as to other firm characteristics.
7. Methodology, empirical strategy and presentation of results
7.1. Computation of basic variables
Log-Returns at time τ for each security i have been computed as:
Ri,τ = ln
(1 +
pi,τ+1 − pi,τpi,τ
), i = 1, . . . , n, τ ∈ (T0, T3]
possible.21See Brown and Warner (1980, 1985) on how to deal with monthly or daily information in connection to
a very short event window.22We adopted this model as a first approximation, though it is clear that the joint distribution of stock
prices is not even asymptotically normal, but rather, it exhibits fat tails and follows a power law (see, e.g.Buchanan, 2008). It might be also possible to try other specifications, for example, instead of definingabnormal returns using the conditional expectation, these could be defined on the basis of the conditionalmedian (i.e. the 0.5 quantile).
14
where pi,τ is the closure price of the stock market security i on day τ , and n = 72 is thenumber of securities traded in the MSE for which we had all the information required.
The MSE market average log-return Rm,τ has been obtained instead as:
Rm,τ =
∑ni=1Ri,τ ×MKi,τ∑n
i=1MKi,τ
, τ ∈ (T0, T3]
where MKi,τ is the market capitalization of security i at time τ .
Figure 3: Market Return weighted by capitalization for the entire 26-week period (Jul, 1922 - Feb, 1923)
In order to build a measure of market capitalization MKi,τ , we have estimated thenumber of shares outstanding for each firm. In particular, we computed:
MKi,τ =SKi
NVi× pi,τ , i = 1, . . . , n, τ ∈ (T0, T3]
where SKi is the share capital of the firm and NVi is the nominal value of the security, forthe benchmark year adopted.
Moreover, dividend payments per share (DpSi) for each firm have been estimated com-bining data on dividends from balance sheet records in IMITA.db and the shares outstandingpreviously obtained:
DpSi =NViSKi
×DIVi, i = 1, . . . , n
15
Figure 4: Unweighted Market Return for the entire 26-week period (Jul, 1922 - Feb, 1923)
16
where DIVi is the dividend payments obtained from the balance sheet for 1922 of firm i.Market capitalization (MKi,τ ) and dividends per share (DpSi), however, are built from
data which has an annual frequency, while our period of interest covers not only July-December, 1922, but also January-February, 1923. As a methdological choice, for datapoints regarding 1923 we considered more appropriate to use data corresponding to theyear 1922. In fact, since we have been considering only the first two months of 1923, usingdata concerning the whole year 1923, given the nature of the present analysis, would haveimplied incorporating into our regressions the effect of events taking place way beyond theevent window of interest.
7.2. Analysis of Abnormal Log-returns and Cumulated Abnormal Log-returns
In concrete terms, to begin with, we considered the following linear probability modelfor each firm during the three-month estimation window (July-October, 1922):
which has been estimated using the OLS method, obtaining αi and γi, for each firm i.We have then estimated abnormal log-returns (AR, hereinafter) for firm i at time τ
during the event-window (T1, T2] by computing:
ARi,τ = Ri,τ − αi − γiRm,τ , τ ∈ (T1, T2] (2)
As rendered clear in equation (2), ARi,τ “is the disturbance term of the market modelcalculated on an out of sample basis” (MacKinlay, 1997, p. 20). In fact, point estimates ofαi and γi for the period (T0, T1] have been used to compute ARs during the event window(T1, T2].23
Using the aggregation rules specified in MacKinlay (1997), we have computed cumulatedabnormal log-returns (CAR, hereinafter) for each firm i during the event window (T1, T2]:
CARi(τ1, τ2) =
τ2∑τ=τ1
ARi,τ , T1 < τ1 ≤ τ2 ≤ T2
as well as ARs and CARs for the group of connnected firms (ncon) and for the rest of firmsoperating in the MSE (nncon):
ARcon,τ =1
ncon
ncon∑i=1
ARi,τ , ARncon,τ =1
nncon
nncon∑i=1
ARi,τ , τ ∈ (T1, T2]
CARcon(τ1, τ2) =
τ2∑τ=τ1
ARcon,τ , CARncon(τ1, τ2) =
τ2∑τ=τ1
ARncon,τ , T1 < τ1 ≤ τ2 ≤ T2
23The estimation window (T0, T1] goes from the beginning of the time period considered (July 1922) toten days before the March on Rome. Thus, ARs have been computed for the period consisting of ten daysbefore to ten days after the event day (October 28th, 1922). There are, however, some missing days inthe dataset due to sundays, holidays and vacations. By ten days we mean ten working days of the stockexchange.
17
A note of caution is in place concerning point estimates for ARs and averaging acrossfirms in a setting where there is total clustering, i.e. all firms share the same chronologicalday for the event study (in our case, the March on Rome). While a multivariate regressionmodel with event-date dummies would be a better model specification (given that otherwisethe test statistic for the significance of the ARs may have poor finite sample properties),24
it has nevertheless been concluded that “results on specification are not radically altered inexperiments where there is clustering in event dates and hence [serial] dependence of theexcess return measures” (Brown and Warner, 1985). Thus, we prefer to keep the frameworkas simple as possible, sticking to our specification in (2).
Tables 4 and 5 report abnormal returns for all connected and non-connected firms, re-spectively, over the time period considered and in the different time subperiods specifiedabove in Section 6. Table 3 summarizes abnormal returns and comulated abnormal returnsfor both connected and non-connected firms during the event window. Results are visualizedin Figures 5 and 6.
Table 3: Abnormal Log-Returns (ARs) and Cumulated Abnormal Log-Returns (CARs) for Connected andOther firms during the event window (Oct-Nov, 1922)
Market modelEvent time ARcon,τ CARcon ARncon,τ CARncon
Source: Own computations based on Il Sole Financial Newspaper and IMITA.db
As can be seen from Figure 5, the evolution of abnormal returns in the days before theMarch on Rome was characterized, for both groups, by strong fluctuations; on day zero,
24For a detailed discussion see MacKinlay (1997, p. 27).
18
connected firms’ abnormal returns peaked, being much higher than those of non-connectedones. In the following days, ARs started fluctuating again, but in a smoother way; ingeneral, and with the exception of day 14, ARs for connected firms were higher than fornon-connected ones.25
Figure 5: Abnormal Returns for Connected and Other firms during the event window
Figure 6: Cumulated Abnormal Returns for Connected and Other firms during the event window
Figure 6 reports cumulated abnormal returns for both connected and non-connectedfirms. As can be seen, the former group was characterized, during the event window, by
25It is worth stressing again that ARs have been computing on the basis of individual firm-level regressionsin which the only independent variable is average market returns. As will be seen in the next Section, thereare other variables, and specifically market capitalization, market beta and industrial sector of activitywhich influences market returns themselves. In particular, some sectors were characterized, during theperiod considered, by above- or below-average returns independently of connectedness of firms.
19
higher CARs than non-connected ones, the gap between the two groups opening exactly incorrespondence of the March on Rome.
Table 4: Summary of Abnormal Log-Returns (ARs) for connected firms
(21-day event window, Oct-Nov, 1922)ARs for the Market model
Ticker Cluster ATECO ARi,pre−event ARi,event−1 ARi,event ARi,post−event ARi
Source: Own computations based on Il Sole Financial Newspaper and IMITA.db
7.3. Cross-sectional effect of connectedness on stock returns
At this point, we estimated the value of Fascist affiliations. Following the methodologicalblueprint of Ferguson and Voth (2008), we asserted the effect on the cross-section of log-returns of political connectedness for firms operating in the MSE between July, 1922 andFebruary, 1923, considering additional controls.
Before proceeding with the analysis of the results, Table 6 provides some descriptivestatistics of the sample used, for connected and non-connected firms, both before and afterthe March on Rome. The sample includes 70 firms, 19 connected and 51 non-connected.
First of all, we can see that market capitalization was higher, in both periods, for non-connected firms; however, the gap reduced after the March. Mean dividend yields are alsohigher for non-connected firms; contrarily to market capitalization, the difference betweenthe two groups deepened after the March. Mean log-returns were higher for connected firmsin both time periods; however, whereas the difference between the two groups was smallerbefore (0.0138 as against 0.0105), such a difference became wider after the March, withaverage log-returns for non-connected firms becoming negative (0.0063 versus -0.0006).
Given the fact that our database includes quite a small number of firms, we used weeklyrather than monthly returns. Moreover, in order to keep as much information as possible, wedecided to compute average weekly returns, i.e. computing for each week average quotations,
20
Table 5: Summary of Abnormal Log-Returns (ARs) for non-connected firms
(21-day event window, Oct-Nov, 1922)ARs for the Market model
Ticker Cluster ATECO ARi,pre−event ARi,event−1 ARi,event ARi,post−event ARi
Connected Other Firms Connected Other FirmsMean Stock Market Capitalization 80978017 83934069 91097587 90430738(in thsd. LIRA)Weight by capitalization in total 0.2644 0.7356 0.2729 0.7271Mean dividend yield 0.0616 0.0650 0.0548 0.0600Mean log-return 0.0138 0.0105 0.0063 -0.0006N 19 51 19 51
Source: Own computations based on Il Sole Financial Newspaper and IMITA.db
and then computing returns based on such averages.26
More specifically, we estimated the following equation:27
log rett = αt + btCL1,t + εt (3)
where CL1,t is a dummy which takes value 1 for firms in the ‘connected’ group.Table 9 shows the results of including into the regressions additional variables: firms-
level market capitalisation and dividend yields, and the market β computed against theperformance of state bonds. Table 10 also includes sectoral dummies.
26As already stressed above, some days are missing due to holidays and vacations. Using average weeklyreturns rather than picking a specific day of the week and computing returns against the same day ofthe following one allows to avoid missing observations. However, in order to check robustness of such achoice, we also computed returns using the alternative method, and then run regressions for each possiblechoice. Results are shown in Table 7. Looking at the top panel of Table, concerning the period beforethe March on Rome, results are qualitatively very close to each other in the different cases considered,i.e. picking as a reference point each single day of the week and weekly averages: the constant is the onlysignificant coefficient, while the one associated to the dummy indicating connected firms is always positivebut not significant. The bottom panel, concerning the period after the March on Rome, displays somemore heterogeneity. More specifically, the intercept is significant for the case of Tuesday, Wednesday andSaturday, while it is not for the other days and for weekly averages. However, the coefficient associated to theconnected-firms dummy is significant and positive in all cases. At the light of these considerations, it seemsreasonable to take weekly averages as our reference point, given the fact that the number of observationsfor the period following the March on Rome is much higher than that which could be obtained by pockingany other day (982 as against 901 which is the highest number of observation for the case of single days).
Table 8 compares the results shown in Table 7, and based on weekly averages, to those obtained byestimating the same equations for the other clusters. In the period before the March on Rome, two clusters(CL3 and CL10) show significantly higher-than-average; however, these two clusters include one single firmeach, i.e. Ceramica Richard Ginori and Cotonificio Val d’Olona, respectively. Parallely, two clusters (CL5
and CL6) significantly lower-than-average log returns; CL6 also consists of one firm only: ManifattureCotoniere Meridionali. As to the period after the March on Rome, CL1 is the only group showing higher-than-average returns; on the contrary, log-returns of clusters CL3, CL8 and CL10 are significantly lower-than-average.
27Standard errors are based on Huber-White heteroscedasticity-consistent estimates and clustered on thelevel of the firm.
22
As can be seen from Table 9, introducing market capitalisation, which has an associatedcoefficient which is significant but equal to zero in both time periods, do not change estimatedcoefficients: belonging to the connected cluster has the effect of increasing log returns by0.7% above the rest of firms. Dividend yields do not significantly affect log-returns before theMarch on Rome, but they (positively) do after; however, the introduction of this variabledo not change estimated coefficient. Also in this case, belonging to CL1 increases log-returns by 0.7% above average. Conversely, the effect of market β is positive and significantbefore the March, increasing log-returns by about 0.6% with respect to the rest, while itis negative, though not significant, after. Moreover, introducing beta into the estimatedequation turns the coefficient associated to CL1 significant before the March too: belongingto the connected groups makes log-return grow about 0.4% more than the remainder offirms. However, this latter effect disappears when sectoral dummies are also introducedinto the estimated equation. In fact in this case, as shown in Table 10, the introductionof market β do not turn the coefficient associated to CL1 significant. However, in all casesconsidered in the right-most panel of the Table—i.e. concerning the period after the Marchon Rome—the effect of belonging to connected firms makes log-returns increase by about0.5%.
Table 10 also reports the coefficients associated to different sectoral dummies, the ref-erence sector being J65, i.e. Financial intermediation.28 It is interesting to have a look atwhich sectors are associated to above-average and below-average log-returns both before andafter the March on Rome.
Looking to column (4), referring to the period before the March and showing estimationsof all variables, three sectors—namely CB13 (Mining of metal ores, 0.0165), DI26 (Manu-facture of other non-metallic mineral products (0.0154)), I60 (Land transport; transport viapipelines 0.0078)—are associated to above-average log-returns with 99% confidence. SectorDK29 (Manufacture of machinery and equipment) shows a significant effect on log-returnsonly when market β is not included among regressors, showing that its effect on log-returns isnot sector-specific but simply associated to the fact that firms in the sector are characterised,on average, by an higher value of β itself. Moreover, two sectors—CA11 (Extraction of crudepetroleum and natural gas -0.0132), DN36 (Manufacture of furniture; manufacturing n.e.c.-0.0074)—are characterised by below-average log-returns.
In general, sectoral effects seem weaker in the period after than before; more specifically,three sectors—(Land transport; transport via pipelines 0.0109) , I61 (Water transport 0.0044),I64 (Post and telecommunications 0.0106)—show above-average log-returns, and one—DI26(Manufacture of other non-metallic mineral products -0.0125)—below-average returns. Inthis case, the introduction among regressors of market β do not alter the significance of thecoefficients at the 99% significance level.
It is worth stressing that while sector I60 is associated to above-average returns bothbefore and after the March, though in the latter case with a smaller effect, firms operat-ing in sector DI26 have significantly above-average returns before the March, which turn
28Table B.11 reports the Code, Label and Description associated to the sector classification adopted byIMITA.db, i.e. Level-3 Ateco (1991) which is the Italian version of Eurostat Nace Rev. 1
23
Table 7: OLS regression, dependent variable: Log-returns. Weekly data. Column week reports resultsbased on weekly averages rather than weekly observations. Standard errors are based on Huber-Whiteheteroscedasticity-consistent estimates and clustered on the level of the firm. CL1 is a dummy which takesvalue 1 for firms belonging to the cluster of connected firms.
Source: Own computations based on Il Sole Financial Newspaper and IMITA.db
significantly below-average after.
24
Tab
le8:
OL
Sre
gres
sion
bas
edon
wee
kly
aver
ages
.D
epen
den
tva
riab
le:
Log-r
etu
rns.
Sta
nd
ard
erro
rsare
base
don
Hu
ber
-Wh
ite
het
eros
ced
asti
city
-con
sist
ent
esti
mat
esan
dcl
ust
ered
on
the
leve
lof
the
firm
.CL1
isa
du
mm
yw
hic
hta
kes
valu
e1
for
firm
sb
elon
gin
gto
the
clu
ster
ofco
nn
ecte
dfi
rms.
mkt
cap
stan
ds
for
mark
etca
pit
ali
zati
on
;div
yiel
ds
for
div
iden
dyie
lds.
Befo
reCL1
CL2
CL3
CL4
CL5
CL6
CL7
CL8
CL9
CL10
Con
stan
t0.
0104
***
0.01
06**
*0.
0110
***
0.0
117***
0.0
122***
0.0
113***
0.0
111***
0.0
114***
0.0
114***
0.0
111***
(0.0
012)
(0.0
011)
(0.0
010)
(0.0
010)
(0.0
011)
(0.0
010)
(0.0
011)
(0.0
010)
(0.0
010)
(0.0
010)
CLi
0.00
340.
0029
0.02
15**
*-0
.0035
-0.0
074***
-0.0
028**
0.0
005
-0.0
067
-0.0
047
0.0
084***
(0.0
020)
(0.0
025)
(0.0
010)
(0.0
031)
(0.0
017)
(0.0
010)
(0.0
024)
(0.0
039)
(0.0
073)
(0.0
010)
N.
Ob
s97
597
597
5975
975
975
975
975
975
975
Aft
er
Con
stan
t-0
.000
70.
0015
0.00
140.0
010
0.0
015
0.0
011
0.0
014
0.0
013
0.0
011
0.0
012
(0.0
009)
(0.0
010)
(0.0
008)
(0.0
009)
(0.0
008)
(0.0
009)
(0.0
010)
(0.0
009)
(0.0
009)
(0.0
009)
CLi
0.00
69**
*-0
.001
9-0
.024
4***
0.0
007
-0.0
028
0.0
016
-0.0
015
-0.0
064*
-0.0
009
-0.0
048***
(0.0
014)
(0.0
018)
(0.0
008)
(0.0
017)
(0.0
032)
(0.0
009)
(0.0
017)
(0.0
027)
(0.0
012)
(0.0
009)
N.
Ob
s98
298
298
2982
982
982
982
982
982
982
Sou
rce:
Ow
nco
mp
uta
tion
sb
ase
don
IlS
ole
Fin
an
cial
New
spap
eran
dIM
ITA
.db
25
Table 9: OLS regression based on weekly averages. Dependent variable: Log-returns. Standard errors arebased on Huber-White heteroscedasticity-consistent estimates and clustered on the level of the firm. CL1
is a dummy which takes value 1 for firms belonging to the cluster of connected firms. mkt cap stands formarket capitalization; div yields for dividend yields.
Source: Own computations based on Il Sole Financial Newspaper and IMITA.db
26
Tab
le10
:O
LS
regr
essi
onb
ased
onw
eekly
aver
ages
.D
epen
den
tva
riab
le:
Log-r
etu
rns.
Sta
nd
ard
erro
rsare
base
don
Hu
ber
-Wh
ite
het
eros
ced
asti
city
-con
sist
ent
esti
mat
esan
dcl
ust
ered
on
the
leve
lof
the
firm
.CL1
isa
du
mm
yw
hic
hta
kes
valu
e1
for
firm
sb
elon
gin
gto
the
clu
ster
ofco
nn
ecte
dfi
rms.
mkt
cap
stan
ds
for
mark
etca
pit
ali
zati
on
;div
yiel
ds
for
div
iden
dyie
lds.
Sec
tora
ld
um
mie
sin
dic
ate
du
sin
gA
TE
CO
clas
sifi
cati
on.
Ref
eren
cese
ctor
:J65
.
Before
After
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Con
stant
0.0
054**
0.0
060*
0.0
036
0.0
011
0.0
002
-0.0
021
-0.0
057*
-0.0
043
(0.0
019)
(0.0
026)
(0.0
027)
(0.0
021)
(0.0
019)
(0.0
020)
(0.0
027)
(0.0
023)
CL1
-0.0
003
-0.0
003
0.0
003
0.0
009
0.0
051**
0.0
049**
0.0
055**
0.0
048*
(0.0
018)
(0.0
018)
(0.0
015)
(0.0
013)
(0.0
017)
(0.0
017)
(0.0
021)
(0.0
019)
mkt
cap
0.0
000
0.0
000
0.0
000
0.0
000
0.0
000
0.0
000
(0.0
000)
(0.0
000)
(0.0
000)
(0.0
000)
(0.0
000)
(0.0
000)
div
yie
lds
0.0
520*
0.0
484*
0.0
748**
0.0
737**
(0.0
242)
(0.0
224)
(0.0
286)
(0.0
272)
β0.0
040**
-0.0
025*
(0.0
014)
(0.0
012)
A01
0.0
074**
0.0
070**
0.0
052*
0.0
055**
0.0
023
0.0
040
0.0
027
0.0
032
(0.0
022)
(0.0
026)
(0.0
024)
(0.0
019)
(0.0
020)
(0.0
023)
(0.0
025)
(0.0
021)
CA
11
-0.0
089***
-0.0
095***
-0.0
125***
-0.0
132***
0.0
013
0.0
035
-0.0
013
-0.0
003
(0.0
019)
(0.0
025)
(0.0
027)
(0.0
017)
(0.0
019)
(0.0
020)
(0.0
027)
(0.0
021)
CB
13
0.0
239***
0.0
235***
0.0
231***
0.0
165***
-0.0
009
0.0
010
0.0
012
0.0
057**
(0.0
019)
(0.0
025)
(0.0
023)
(0.0
022)
(0.0
019)
(0.0
019)
(0.0
020)
(0.0
020)
DA
15
0.0
115**
0.0
111**
0.0
094*
0.0
082*
0.0
023
0.0
036
0.0
022
0.0
036
(0.0
038)
(0.0
039)
(0.0
043)
(0.0
040)
(0.0
023)
(0.0
024)
(0.0
025)
(0.0
022)
DB
17
0.0
095***
0.0
091***
0.0
070*
0.0
036
-0.0
010
0.0
008
-0.0
014
0.0
013
(0.0
023)
(0.0
027)
(0.0
028)
(0.0
020)
(0.0
022)
(0.0
022)
(0.0
024)
(0.0
022)
DG
24
0.0
116*
0.0
113*
0.0
125*
0.0
102*
0.0
043
0.0
051
0.0
073
0.0
087*
(0.0
056)
(0.0
056)
(0.0
054)
(0.0
043)
(0.0
033)
(0.0
035)
(0.0
045)
(0.0
043)
DI2
60.0
271***
0.0
266***
0.0
271***
0.0
154***
-0.0
231***
-0.0
211***
-0.0
203***
-0.0
125***
(0.0
019)
(0.0
025)
(0.0
024)
(0.0
040)
(0.0
019)
(0.0
019)
(0.0
022)
(0.0
033)
DJ27
0.0
093*
0.0
090
0.0
112*
0.0
097*
-0.0
048
-0.0
036
-0.0
004
0.0
007
(0.0
047)
(0.0
049)
(0.0
049)
(0.0
039)
(0.0
039)
(0.0
041)
(0.0
046)
(0.0
037)
DJ28
0.0
023
0.0
018
0.0
006
0.0
030
0.0
018
0.0
040*
0.0
026
0.0
015
(0.0
019)
(0.0
025)
(0.0
024)
(0.0
019)
(0.0
019)
(0.0
020)
(0.0
022)
(0.0
016)
DK
29
0.0
150***
0.0
144***
0.0
162***
0.0
058
-0.0
058**
-0.0
034
-0.0
004
0.0
066
(0.0
022)
(0.0
027)
(0.0
025)
(0.0
035)
(0.0
020)
(0.0
025)
(0.0
033)
(0.0
039)
DM
34
0.0
027
0.0
027
0.0
032
-0.0
004
-0.0
104
-0.0
103
-0.0
092
-0.0
068
(0.0
025)
(0.0
024)
(0.0
018)
(0.0
019)
(0.0
073)
(0.0
059)
(0.0
053)
(0.0
046)
DM
35
0.0
009
0.0
005
-0.0
014
-0.0
014
-0.0
001
0.0
015
-0.0
008
-0.0
003
(0.0
031)
(0.0
034)
(0.0
034)
(0.0
034)
(0.0
020)
(0.0
018)
(0.0
023)
(0.0
016)
DN
36
-0.0
059**
-0.0
064*
-0.0
087***
-0.0
074***
-0.0
018
0.0
003
-0.0
035
-0.0
038*
(0.0
019)
(0.0
025)
(0.0
025)
(0.0
018)
(0.0
019)
(0.0
020)
(0.0
025)
(0.0
018)
E40
0.0
000
-0.0
002
-0.0
009
0.0
001
0.0
017
0.0
027
0.0
022
0.0
020
(0.0
022)
(0.0
024)
(0.0
023)
(0.0
016)
(0.0
024)
(0.0
020)
(0.0
021)
(0.0
015)
G52
-0.0
064**
-0.0
066**
-0.0
040
-0.0
013
-0.0
010
0.0
002
0.0
040
0.0
024
(0.0
019)
(0.0
022)
(0.0
024)
(0.0
020)
(0.0
019)
(0.0
016)
(0.0
023)
(0.0
021)
I60
0.0
152***
0.0
147***
0.0
119***
0.0
078***
0.0
071*
0.0
092*
0.0
077***
0.0
109***
(0.0
023)
(0.0
028)
(0.0
031)
(0.0
016)
(0.0
035)
(0.0
036)
(0.0
022)
(0.0
021)
I61
-0.0
046
0.0
002
-0.0
002
0.0
009
0.0
007
0.0
050***
0.0
048**
0.0
044***
(0.0
040)
(0.0
019)
(0.0
018)
(0.0
013)
(0.0
035)
(0.0
014)
(0.0
015)
(0.0
011)
I64
0.0
040*
0.0
034
0.0
058*
0.0
043*
0.0
036
0.0
059**
0.0
095***
0.0
106***
(0.0
019)
(0.0
026)
(0.0
027)
(0.0
018)
(0.0
019)
(0.0
020)
(0.0
027)
(0.0
019)
N.
Ob
s975
962
962
962
982
969
969
969
Sou
rce:
Ow
nco
mp
uta
tion
sb
ase
don
IlS
ole
Fin
an
cial
New
spap
eran
dIM
ITA
.db
27
Appendix A. A historiographic analysis of prominent individuals within thecluster of connected firms
Appendix A.1. Lodovico Mazzotti Biancinelli
Lodovico Mazzotti Biancinelli’s entrepreneurial activities began in 1900, when he and apartner founded a company that operated as selling agent in connection with COMIT. In1904, having wound up this company, Mazzotti Biancinelli became fiduciary of the COMITat the MSE. In 1909 he was nominated president of the MSE, an office he retained until1920.
In these years, Mazzotti Biancinelli also commenced his activities of stock gamblingand also land reclamation in some provinces of Northern Italy. Initially, his main businessrelated to the sugar industry. Between 1910 and 1920 his name appeared on the boards ofalmost a dozen companies in this sector, together with other well-known stock gamblers.This circumstance leads us that sometimes their profits were due to the watering of stock(Romani, 2008).
In the years preceding World War One, Mazzotti Biancinelli started to expand his rangeof business. From 1910 he was on the board of Ilva and, from 1914, on that of the bankCredito Agrario Bresciano.
In August 1911, in the face of financial difficulties besetting the entire metallurgic sector,the government pushed six of the major companies within the sector (Elba, Ilva, Savona,Ligure Metallurgica, Ferriere Italiane and Alti Forni di Piombino) to enter into an agreementfor a unique management of their factories. This was the so-called Consorzio Ilva. The ideawas that Ilva, headed by Attilio Odero, would manage the factories of the pool for a periodof eleven years and six months while each company would have a stake in operating profitsaccording to the productive efficiency of their respective factories. The capital of this poolamounted to 130 million lire, of which 96 million lire was financed by the Bank of Italy,COMIT and Credito Italiano (Cerioni, 2001). The eight factories of the group manufacturedthe entire Italian production of iron ore and cast iron and 58% of Italian steel.
In 1916 Mazzotti Biancinelli became vice-president of the Credito Agrario Bresciano.Under his management, the main business of the bank shifted rapidly from the agriculturalto the industrial sector. Specifically, the bank started to finance those companies in whichMazzotti Biancinelli was interested: SADE, Isotta Fraschini, Societa elettrica Milano, Co-tonificio Veneziano, Societa elettrica Riviera di Ponente, Fabbriche italiane materie colorantiBonelli, Cotonificio Turati, Officine meccaniche, and many others. Because of this kind ofoperation, in 1919 Mazzotti Biancinelli was compelled to resign from the bank.
In 1918 Mazzotti Biancinelli helped Max Bondi in his takeover of Ilva. Prior to theestablishment of Consorzio Ilva, Bondi headed the Alti Forni di Piombino and was on theboards of Eridania, the Societa romana per la fabbricazione dello zucchero, Elba and manyothers. Once he gained majority control of Ilva, Bondi changed the company name from AltiForni di Piombino to Ilva altiforni e accierie d’Italia. Under this new name now operatedthe old Ilva, Savona, Ligure Metallurgica, and Ferriere Italiane, while Elba remained anindependent company (Carparelli, 1978). Bondi’s management was disastrous and, as wehave seen in section 3, Mussolini’s first government, breaking its laissez-faire economic policy,
28
rescued Ilva by reducing the amount of its debt to the state but allowing COMIT and CreditoItaliano to gain majority control of the company.
Appendix A.2. Giuseppe Da Zara
Giuseppe Da Zara started his career during the 1880s in the insurance sector. While stillyoung, he gained a seat on the board of Assicurazioni Generali, one of most ancient andprestigious Italian companies in the insurance business. Between 1901 and 1909 Da Zarawas deputy director of this company, and then (until 1922) director. In 1911 Da Zara wasa leading figure in the campaign against the government’s plan to establish a monopoly oflife assurances.
At the same time, Da Zara expanded his interests to include the banking sector (from1900 he was on the board of the Societa Bancaria Milanese, a bank that aimed to counterthe predominance of COMIT and Credito Italiano at the national level) as well as themetallurgic sector. From the early 1900s, he was on the board of Terni and then on that ofElba and Ilva during the years of Consorzio Ilva.
Da Zara’s connection to Giuseppe Volpi da Misurata continued for many years. Theirpoints of contact ranged from SADE to a group of companies operated in Veneto, such asSocieta veneta per la costruzione ed esercizio di ferrovie secondarie (a railway company), theBanca veneta di depositi e conti correnti (a bank), and the Societa veneziana di navigazionea vapore (a steam navigation company) (Segreto, 1987).
After the end of the World War One, Da Zara was implicated in the financial collapse ofBanca Italiana di Sconto because of his position in Societa Bancaria Milanese. At the endof the trial of 1922-1923 he was acquitted of any wrong doing. He died suddenly in 1923.
Appendix A.3. Giuseppe Volpi di Misurata
At the end of nineteenth century, Giuseppe Volpi was operating in the Balkans as adealer in agricultural products, insurance agent and mining agent. In 1900 an encounterwith Giuseppe Toeplitz, then manager of the branch of COMIT in Venice, triggered a careerin finance. In 1915 Toeplitz was nominated member of the board of COMIT and in 1917 hebecame chief executive of the bank (a post he occupied until 1933).
Volpi’s early entrepreneurial activities were financed by Toeplitz. He banked on a web ofpersonal relationships in the political and financial circles of Venice (Da Zara, for example,represented a link with Venetian Jewish finance). Broadly, Volpi and this group aimed forthe Balkan market, a plan of financial expansion shared by Toeplitz for COMIT (Tamborra,1974).
In 1900 Volpi, with the financial aid of this group, founded the Societa Italiana perl’utilizzazione delle forze idrauliche del Veneto (an electrical company) and, in 1905, SADE.In both cases, the financial support of COMIT was crucial: in the first case COMIT sub-scribed 18% of the initial capital; in the other 16%. In the following years, SADE progres-sively became the primary holding of the group and at the end of World War One SADEwas the third largest electrical Italian company for capital capacity (Petri and Reberschack,1993).
29
In 1906 Volpi and his group extended their interest to the hotel industry with the estab-lishment of the Compagnia Italiana Grandi Alberghi (hereafter, CIGA); once again COMITshared nearly 20% in the initial capital of the society.
During these years, COMIT operated in the Balkans with the Regia interessata deitabacchi and the Sindacato italo-montenegrino, which were both founded in 1903, and theCompagnia di Antivari (1905). In 1907, COMIT founded the Societa Commerciale d’Oriente,which represented a sort of Montenegrin branch of the bank. The strategist of these oper-ations was undoubtedly Volpi who, over the years, had stabilized his economic relations inthe Balkans and expanded his economic interests to Turkey.
At the end of World War One, Volpi acted as government delegate on several interna-tional commissions. In December 1920 he obtained the aristocratic title of count because ofservices rendered to Italy and, in June 1921, he became governor of Libya, which countryhe succeeded in stabilizing. Thanks to this political (and military) success he acquired 2000hectares of land close to Misurata and in recognition of his efforts was allowed to add “diMisurata” to his aristocratic title.
Historical research has recently established that in the years 1920-1922 Volpi financedsuccessively the fascist movement of Venice and then the local PNF (Piva, 1977; Albanese,2001).
From 1921 to 1922 Volpi held the office of president or vice-president of twenty societies,among them three electrical companies (SADE, Societa Italiana per l’utilizzazione delle forzeidrauliche del Veneto and S.I.V.), financial companies such, as for example, the CreditoIndustriale, and industrial companies such as CIGA. Moreover, he was on the boards of 46societies, among them COMIT, Assicurazioni Generali, and Unione Esercizi Elettrici.
On July 13th 1925 Volpi replaced De Stefani as the Minister of Finance.
Appendix A.4. Emilio Bruzzone
Bruzzone headed the Ligure lombarda, a Genoa sugar business. As we saw in section 3,he belonged to the group who first financed Mussolini’s Popolo d’Italia.
The Italian sugar industry was mainly located in Genoa. The first development of thissector in Italy was linked with the emergence of the refining industry. In the first half of thenineteenth century Italian production was modest and from 1870 almost 50% of importedsugar arrived in Genoa. This circumstance favoured the birth of the sugar refining industryin Genoa. The introduction of a duty on imported sugar in the 1870s gave the impetus tonational production and, effectively, to those Genoese industrialists who already operatedin the sugar refining sector. In 1914, 13 out of 26 companies operating in the sugar sectorhad their seats in Genoa.
The sector was characterized by a strong concentration and the companies, which in1903 established the Unione Zuccheri, acted as a cartel. By means of cross shareholdings,Ligure lombarda, together with Eridania, held the majority control of 16 other companiesand, through them, of 30 factories. Ligure lombarda and Eridania controlled 61% of Italianplants and 53% of national production. The Societa italiana per l’industria dello zuccheroindigeno was up there with them. This latter society, which was controlled by ErasmoPiaggio and Emilio Maraini, together with Fabbrica Lendinarese, which was exclusively
30
owned by Maraini, had a nominal capital of 22.57 million lire, managed ten factories with anannual production of 47,000 tons of sugar that amounted to 21% of the national production(Tonizzi, 2001).
Appendix A.5. Benedetto Acquarone
Giacomo Acquarone was one of the founders of Eridania in Genoa in 1899; he remainedon the board of the company until 1902. In May 1922 his son, Benedetto Acquarone, whohad inherited his father’s shares, entered onto the board of Eridania, on which was alreadysitting his uncle, Serafino Cevasco, as managing director. Acquarone gained work experienceat the San Giorgio, an electrical-mechanical industry in Genoa (Bianchi Tonizzi, 1988).
Acquarone and Cevasco implemented the vertical integration of sugar-beet productionand sugar refining and production and its by-products (such as leaven, manna sugar andalcohol). Moreover, they worked together on some Italian mechanical factories in order tostart up production of the industrial equipment necessary for refining and producing sugaron a large scale.
In the first months of 1923 Aldo Finzi,29 vice-minister for Home Affairs, got in touch withrepresentatives of industry in order to ask for financial support for the founding of a newFascist newspaper in Rome, Il Corriere d’Italia. On April 14th 1923, with funds obtainedfrom Ilva, Piaggio, Eridania, Ansaldo, and Odero, Finzi established the publishing houseLa vita d’Italia editrice, which published Il Corriere d’Italia (Canali, 2009, 78-85). Finzinot only held a bundle of shares in the company but managed the newspaper with FilippoFilippelli.30
On November 20th 1930, Acquarone and Cevasco realized the merger of Eridania withLigure lombarda. The new company, named Eridania zuccherifici nazionali, controlled Dis-tillerie italiane and Saccarifere lombarda, and managed 21 sugar refineries in addition to theplants for the manufacturing of the by-products. Its social capital amounted to 120 millionlire and the value of plants amounted to 400 million lire.
Appendix B. Statistical Companion
29Finzi joined the Fascist Action Squads in 1920. In 1921 he was elected deputy and he was one of themost resolute opponent to Patto di Pacificazione with the Socialist Party. On June 13th 1922, together withother Fascists he attacked the communist deputy, Francesco Misiano. He was implicate in the Matteottiassassination and Mussolini compelled him to resign, even though he was not brought to trial (Canali, 2004).
30The Corriere d’Italia closed after Matteotti’s assassination because its director, Filippelli, was arrestedfor abetment in the kidnapping and assassination of Matteotti (Canali, 2004).
Italian version of EUROSTAT NACE Rev. 1 (1990)Code Label DescriptionA01 Agriculture AgricultureA02 Forestry ForestryB05 Fishing Fishing
CA10 Coal-lignite-peat Mining of coal and lignite; extraction of peatCA11 Petroleum-Gas Extraction of crude petroleum and natural gasCB13 Metal-ores Mining of metal oresCB14 Other-mining Other mining and quarryingDA15 Food-beverages Manufacture of food products and beveragesDB17 Textiles Manufacture of textilesDB18 Wearing-apparel Manufacture of wearing apparel; dressing and dyeing of furDC19 Leather-footwear Tanning and dressing of leather; manufacture of luggage, handbags,
saddlery, harness and footwearDD20 Wood Manufacture of wood and of products of wood and corkDE21 Pulp-paper Manufacture of pulp, paper and paper productsDE22 Publishing-printing Publishing, printing and reproduction of recorded mediaDF23 Refined-petroleum Manufacture of coke, refined petroleum productsDG24 Chemicals Manufacture of chemicals and chemical productsDH25 Rubber-plastic Manufacture of rubber and plastic productsDI26 Non-metallic-mineral Manufacture of other non-metallic mineral productsDJ27 Basic-metals Manufacture of basic metalsDJ28 Fabricated-metals Manufacture of fabricated metal productsDK29 Machinery-equipment Manufacture of machinery and equipmentDL30 Office-machinery Manufacture of office machinery and computersDL31 Electrical-machinery Manufacture of electrical machinery and apparatusDL32 Communication-equip Manufacture of radio, television and communication equipment and
apparatusDL33 Precision-equip Manufacture of medical, precision and optical instruments, watches
and clocksDM34 Motor-vehicles Manufacture of motor vehicles, trailers and semi-trailersDM35 Transport-equip Manufacture of other transport equipmentDN36 Manufacturing-nec Manufacture of furniture; manufacturing n.e.c.E40 Electricity-gas Electricity, gas, steam and hot water supplyE41 Water Collection, purification and distribution of waterF45 Construction ConstructionG50 Repair-fuel Sale, maintenance and repair of motor vehicles and motorcycles; retail
sale of automotive fuelG51 Wholesale-trade Wholesale trade and commission tradeG52 Retail-trade Retail tradeH55 Hotels-restaurants Hotels and restaurantsI60 Land-transport Land transport; transport via pipelinesI61 Water-transport Water transportI62 Air-transport Air transportI63 Supporting-transport Supporting and auxiliary transport activitiesI64 Post-Telecomm Post and telecommunicationsJ65 Finance Financial intermediationJ66 Insurance Insurance and pension fundingK70 Real-estate Real estate activitiesK71 Renting-machinery Renting of machinery and equipmentK74 Other-business-acts Other business activitiesM80 Education EducationN85 Health Health and social workO92 Recreation-culture Recreational, cultural and sporting activitiesO93 Other-services Other service activities
Agrawal, A. and Knoeber, C. (2001). Do Some Outside Directors Play a Political Role? Journal of Lawand Economics, 44:179–198.
Albanese, G. (2001). Alle origini del fascismo. La violenza politica a Venezia 1919-1922. Il Poligrafo, Padova.Amatori, F. and Brioschi, F. (1997). Le grandi imprese private: famiglie e coalizioni. In Barca, F., editor,
Storia del capitalismo italiano dal dopoguerra ad oggi, pages 117–153. Donzelli, Roma.Amatori, F., Bugamelli, M., and Colli, A. (2011). Italian Firms in History: Size, Technology and En-
trepreneurship. Banca d’Italia, Quaderni di Storia Economica (Economic History Working Papers), 13.Amatori, F. and Colli, A. (1999). Impresa e industria in Italia dall’Unita ad oggi. Venezia, Marsilio.Asso, P. F. and De Cecco, M. (1994). Storia del Crediop: tra credito speciale e finanza pubblica 1920-1960.
Laterza, Bari.Associazione fra le societa italiane per azioni, editor (1928). Notizie statistiche sulle societa italiane per
azioni 1928. Roma.Associazione fra le societa italiane per azioni, editor (1937). Societa italiane per azioni. Notizie statistiche
1937. Roma.Barbone, D. (2003). Ancora sulle elargizioni della COMIT a Fascisti (1919-1930). Quaderni di Storia,
58(2):259–279.Belloni, E. (2011). Confindustria e lo sviluppo economico italiano. Gino Olivetti tra Giolitti e Mussolini. Il
Mulino, Bologna.Bianchi Tonizzi, E. (1988). Acquarone, Benedetto. In Dizionario Biografico degli Italiani, volume 34. Istituto
dell’Enciclopedia Italiana Treccani, Roma.Bonelli, F. (1971). La crisi del 1907. Una tappa dello sviluppo industriale in Italia. Fondazione L. Einaudi,
Turin.Boubakri, N., Guedhami, O., Misha, D., and Saffar, W. (2012). Political Connections and the Cost of Equity
Capital. Journal of Corporate Finance, 18:541–559.Brown, S. J. and Warner, J. B. (1980). Measuring security price performance. Journal of financial Economics,
8(3):205–258.Brown, S. J. and Warner, J. B. (1985). Using daily stock returns: The case of event studies. Journal of
financial economics, 14(1):3–31.Buchanan, M. (2008). The social atom: Why the rich get richer, cheaters get caught, and your neighbor
usually looks like you. Bloomsbury USA.Bunkanwanicha, P. and Wiwattantang, Y. (2009). Big Business Owners in Politics. The Review of Financial
Studies, 22(6):2133–2166.Campbell, J., Lo, A., and MacKinlay, C. (1997). The Econometrics of Financial Markets. Princeton
University Press, Princeton, NJ.Canali, M. (2004). Il delitto Matteotti. Il Mulino, Bologna.Canali, M. (2009). The matteotti murder and the origins of mussolini’s totalitarian fascist regime in italy.
Journal of Modern Italian Studies, 14(1):143–167.Civilize, S., Wongchoti, U., and Young, M. (2015). Political Connections and Stock Returns: A Longitudinal
Study. The Financial Review, 50:89–119.Colli, A. (2006). L’impresa italiana nel Novecento (review). Enterprise & Society, 7(4):829–832.Colli, A. and Vasta, M. (2010). Introduction: forms of enterprise in 20th century Italy. In Colli, A. and
Vasta, M., editors, Forms of Enterprise in 20th Century Italy, pages 1–24. Edward Elgar, Cheltenham,UK.
Consob (2011). Dall’Unita ai giorni nostri: 150 anni di borsa in Italia.Credito Italiano, editor (1919). Notizie statistiche sulle principali societa italiane per azioni 1918, 2 vol.
Roma.Credito Italiano, editor (1921). Societa italiane per azioni. Notizie statistiche 1920, 4 vol. Milano.Credito Italiano, editor (1923). Societa italiane per azioni. Notizie statistiche 1922. Roma.Credito Italiano, editor (1926). Societa italiane per azioni. Notizie statistiche 1925. Roma.
39
De Cecco, M., editor (1993). L’Italia e il Sistema Finanziario Internazionale, 1919-1936. Laterza, Bari-Roma.
De Felice, R. (1966a). Mussolini il Fascista, 2 vols. Einaudi, Turin.De Felice, R. (1966b). Mussolini il rivoluzionario. Einaudi, Turin.De Luca, G., editor (2002). Le societa quotate alla borsa valori di Milano dal 1861 al 2000 — Profili storici
e titoli azionari. Libri Scheiwiller.Fabre, G. (2003). Mussolini e le sovvenzioni della COMIT. Quaderni di Storia, 57(1):281–299.Faccio, M., Masulis, R. W., and McConnell, J. J. (2006). Political Connections and Corporate Bailouts.
The Journal of Finance, LXI(6):2597–2635.Ferguson, T. (1995). Golden Rule: The Investment Theory of Party Competition and the Logic of Money-
Driven Political System. Chicago University Press, Chicago.Ferguson, T. and Voth, H.-J. (2008). Betting on Hitler — The Value of Poltical Connections in Nazi
Germany. Quarterly Journal of Economics, 123:101–137.Fisman, R. (2001). Estimating the value of political connections. The American Economic Review,
91(4):1095–1102.Giannetti, R. and Vasta, M., editors (2006). Evolution of Italian enterprises in the 20th century. Physica-
Verlag, Heidelberg.Guerin, D. (1956). Fascismo e gran capitale. Schwarz Editore, Milano.Johnson, S. and Mitton, T. (2003). Cronysm and Capital Controls: Evidence from Malaysia. Journal of
Financial Economics, 67:351–382.Khwaja, A. L. and Mian, A. (2005). Do Lenders Favor Politically Connected Firms? Rent Provision in an
Emerging Financial Market. The Quarterly Journal of Economics, pages 1371–1411.Lamoreaux, N. R., Raff, D. M. G., and Temin, P. (2007). Economic Theory and Business History. In Jones,
G. and Zeitlin, J., editors, The Oxford Handbook of Business History, pages 37–66. Oxford UniversityPress, Oxford.
Leicht, E. and Newman, M. (2008). Community structure in directed networks. Phys. Rev. Lett., 100.Li, H., Meng, L., Wang, Q., and Zhou, L. (2008). Political Connections, Financing and Firm Performance:
Evidence from Chinese Private Firms. Journal of Development Economics, 87:283–299.MacKinlay, A. C. (1997). Event Studies in Economics and Finance. Journal of Economic Literature,
35(1):13–39.Melograni, P. (1972). Gli industriali e Mussolini. Rapporti tra Confindustria e fascismo dal 1919 al 1929.
Longanesi, Milano.Ministero del Tesoro (1999). Il debito pubblico in italia 1861-1987. Technical report, Direzione Generale del
Debito Pubblico – Relazione del Direttore Generale alla Commissione Parlamentare di Vigilanza.Newman, M. (2006a). Finding community structure in networks using the eigenvectors of matrices.
Phys. Rev. E, 74(3).Newman, M. (2006b). Modularity and community structure in networks. Proc. Natl. Acad. Sci. USA,
103:8577–8582.Petri, R. and Reberschack, M. (1993). La SADE di Giuseppe Volpi e la nuova ‘Venezia’ industriale. In
De Rosa, L., editor, Storia dell’industria elettrica in Italia, pages 317–346. Roma-Bari, Laterza.Piva, F. (1977). Lotte contadine e origini del fascismo. Marsilio, Venezia.Roberts, B. (1990). A Dead Senator Tells no Lies: Seniority and the Distribution of Federal Benefits.
American Journal of Political Science, 34:31–58.Romani, M. (2008). Mazzotti Biancinelli, L. In Dizionario Biografico degli Italiani, volume 72. Istituto
dell’Enciclopedia Italiana Treccani, Roma.Rossi, E. (1955). I padroni del vapore. Laterza, Bari-Roma.Sarti, R. (1977). Fascismo e grande industria: 1919-1940. Mozzi Editore, Milano.Segreto, L. (1987). Da Zara, Giuseppe. In Dizionario Biografico degli Italiani, volume 33. Istituto
dell’Enciclopedia Italiana Treccani, Roma.Shleifer, A. and Vishny, R. W. (1994). Politicians and Firms. The Quarterly Journal of Economics,
109(4):995–1025.
40
Tamborra, A. (1974). The rise of Italian industry and the Balcans (1900-1914). The Journal of EuropeanEconomic History, 3(1):87–120.
Toniolo, G. (1980). L’economia dell’Italia fascista. Laterza, Bari-Roma.Tonizzi, M. E. (2001). L’industria dello zucchero. Franco Angeli, Milano.Vivarelli, R. (1992). Storia delle origini del fascismo, 3 vols. Il Mulino, Bologna.
